{"id":"W4400261110","doi":"10.2196/57674","title":"Data Set and Benchmark (MedGPTEval) to Evaluate Responses From Large Language Models in Medicine: Evaluation Development and Validation","year":2024,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":22,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Computer science; Benchmark (surveying); Set (abstract data type); Data set; Data mining; Natural language processing; Artificial intelligence; Data science; Programming language; Cartography","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004458348,0.0001347998,0.000239868,0.0002952543,0.00006855604,0.00004895121,0.0001277511,0.0001669339,0.0004150644],"category_scores_gemma":[0.001760633,0.0001043968,0.000009776741,0.0003230244,0.00006787292,0.0004657402,0.0001545742,0.0002893604,0.00005345083],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001491538,"about_ca_system_score_gemma":0.0008938578,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002497216,"about_ca_topic_score_gemma":0.0002680053,"domain_scores_codex":[0.9974706,0.0001183658,0.0008440252,0.0002195848,0.00111204,0.0002354443],"domain_scores_gemma":[0.9985639,0.0005431537,0.00006243486,0.0003482482,0.0001273565,0.0003549148],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001580923,0.00006307794,0.002505823,0.0006967564,0.00004083995,0.00002047231,0.2347891,0.00001429541,0.00006356334,0.0003527967,0.009261917,0.7520332],"study_design_scores_gemma":[0.0004544502,0.0002023539,0.006390984,0.003135089,0.0001313115,0.00004573167,0.05449455,0.9101601,0.0008584945,0.002911261,0.02096035,0.0002553518],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9884032,0.001527174,0.002759361,0.005478229,0.000317826,0.0008987091,0.00005205137,0.00004997691,0.0005134641],"genre_scores_gemma":[0.991798,0.0003139735,0.003706258,0.001935293,0.0002782005,0.0001262421,0.001764137,0.0000130875,0.00006476276],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9101458,"threshold_uncertainty_score":0.4544664,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2744553495825037,"score_gpt":0.5143539930829046,"score_spread":0.239898643500401,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}